Effect of a participatory whole-systems approach on mortality in children younger than 5 years in Jigawa state, Nigeria (INSPIRING trial): a community-based, parallel-arm, pragmatic, cluster randomised controlled trial and concurrent mixed-methods process evaluation
Carina King,Rochelle Ann Burgess,Ayobami A Bakare,Funmilayo Shittu,Julius Salako,Damola Bakare,Obioma C Uchendu,Agnese Iuliano,Nehla Djellouli,Adamu Isah,Ibrahim Haruna,Samy Ahmar,Tahlil Ahmed,Paula Valentine,Temitayo Folorunso Olowookere,Matthew MacCalla,Hamish R Graham,Eric D McCollum,James Beard,Adegoke G Falade,Tim Colbourn,INSPIRING Project Consortium,Ayobami Adebayo Bakare,Christine Cassar,Hamish G Graham,Abdullahi Magama,Omotayo Olojede,Adams Osebi,Ibrahim Seriki,Abiodun Sogbesan,Obioma Uchendu
DOI: https://doi.org/10.1016/S2214-109X(24)00369-3
Abstract:Background: In 2019, Nigeria reported the highest mortality rate in children younger than 5 years globally. We aimed to assess a whole-systems approach to improving child mortality in northern Nigeria. Methods: We conducted a community-based, parallel-arm, pragmatic, cluster randomised controlled trial in Kiyawa local government area, Jigawa state, Nigeria, and a concurrent mixed-methods process evaluation using ethnography and quantitative implementation monitoring. Trial clusters were population catchment areas of 32 government primary health-care facilities. Compounds were randomly sampled, proportional to cluster size, and all women aged 16-49 years and children younger than 5 years who were permanent residents were eligible for inclusion and recruited as the evaluation population. Children younger than 7 days were recruited but excluded from analysis. Evaluation clusters were allocated to intervention or control via simple randomisation with a 1:1 ratio. Cluster names were written on paper, folded, and placed in a container by community representatives. Different community representatives took out names one by one, with the first half assigned to receive the intervention. The intervention consisted of three components: participatory learning and action (PLA) groups for men and women (including compound heads [ie, the member of the compound that residents deemed most senior]), partnership defined quality scorecard (PDQS), and health-care worker capacity building; it was delivered from March 1, 2021, to Dec 31, 2022. We could not mask participants, field staff, or intervention-delivery staff to cluster allocation but baseline, endline, and follow-up data excluded information on cluster allocation. PLA groups involved separate groups of up to 25 men or women from all villages in the intervention clusters. The primary outcome was all-cause mortality in children aged 7 days to 59 months between Oct 1, 2021, and Sept 20, 2022, referred to as the evaluation period. The trial was prospectively registered (ISRCTN 39213655) and the protocol has been published. Findings: We recruited 3800 compounds at baseline, with 12 893 children contributing to analysis of the primary outcome (7316 [56·8%] of 12 893 in the intervention group and 5577 [43·3%] in the control group). 6617 (51·3%) of 12 893 children were male, 6275 (48·7%) were female, and one (<0·1%) child had missing sex data. Sampled compounds randomly came from 388 (91·3%) of 425 villages in the 32 clusters. We conducted verbal autopsies for 1182 deaths, of which 369 (31·2%) were children aged 7 days to 59 months during the evaluation period. Of these 369, 91 (24·7%) were classified as pneumonia deaths. Children contributed a median 361 days (IQR 236-365) to the analysis, with 369 (2·9%) of 12 893 children censored on their date of death, 1545 (12·0%) on their 5th birthday, and 3392 (26·3%) on the date of the most recent follow-up in which their residence or survival status was known. We found no significant decrease in all-cause mortality (hazard ratio 0·95, 95% CI 0·68-1·33; p=0·79) or suspected pneumonia mortality (0·79, 0·43-1·46; p=0·46) in the intervention group. The process evaluation showed low coverage and issues in reach of the intervention, but qualitative data highlighted mechanisms for positive effects on health and relationships. Interpretation: Our intervention did not affect mortality. However, due to the high child mortality in this region, further efforts should be made to adapt our participatory whole-systems approach to use communities of action within compounds. Funding: GSK and Save the Children UK. Translation: For the Hausa translation of the abstract see Supplementary Materials section.